Journal of Wildlife Diseases, 41(2), 2005, pp. 342–353 ᭧ Wildlife Disease Association 2005

CHARACTERIZATION OF -LIKE INFECTING ENDEMIC BIRDS IN THE GALAÂ PAGOS ISLANDS

Teresa Thiel,1 Noah K. Whiteman,1 Ana TirapeÂ,2,3 Maria Ines Baquero,2,4 Virna CedenÄo,2,3,4 Tim Walsh,5,6,7 Gustavo JimeÂnez UzcaÂtegui,6 and Patricia G. Parker1,5 1 Department of Biology, University of Missouri±St. Louis, Missouri 63121, USA 2 Laboratory of Fabricio Valverde, GalaÂpagos National Park, Isla Santa Cruz, GalaÂpagos, Ecuador 3 Concepto Azul, Guayaquil, Ecuador 4 Biotechnology Program, University of Guayaquil, Guayaquil, Ecuador 5 Charles Darwin Research Station, Puerto Ayora, Isla Santa Cruz, GalaÂpagos, Ecuador 6 Current address: Washington State University, Washington Animal Disease Diagnostic Laboratory, Pullman, Washington 99165-2037, USA 7 Corresponding author (email: [email protected])

ABSTRACT: The presence of avian pox in endemic birds in the Gala´pagos Islands has led to concern that the health of these birds may be threatened by introduction by domestic birds. We describe here a simple polymerase chain reaction–based method for identification and discrimination of avipoxvirus strains similar to the fowlpox or canarypox viruses. This method, in conjunction with DNA sequencing of two polymerase chain reaction–amplified loci totaling about 800 bp, was used to identify two avipoxvirus strains, Gal1 and Gal2, in pox lesions from yellow warblers (Dendroica petechia), finches (Geospiza spp.), and Gala´pagos mockingbirds (Nesomimus parvulus) from the inhabited islands of Santa Cruz and Isabela. Both strains were found in all three taxa, and sequences from both strains were less than 5% different from each other and from canarypox . In contrast, chickens in Gala´pagos were infected with a virus that appears to be identical in sequence to the characterized fowlpox virus and about 30% dif- ferent from the canarypox/Gala´pagos group viruses in the regions sequenced. These results in- dicate the presence of canarypox-like viruses in endemic passerine birds that are distinct from the fowlpox virus infecting chickens on Gala´pagos. Alignment of the sequence of a 5.9-kb region of the revealed that sequence identities among Gal1, Gal2, and canarypox viruses were clustered in discrete regions. This indicates that recombination between poxvirus strains in com- bination with mutation led to the canarypox-like viruses that are now prevalent in the Gala´pagos. Key words: Avian pox, avipoxvirus, canary pox, fowlpox, Gala´pagos.

INTRODUCTION in Hawaii (Wikelski et al., 2004), there is The Gala´pagos Islands are volcanic in increasing concern about the introduction origin (Christie et al., 1992; White et al., of avian diseases that could result in ex- 1993) and located on the equator almost tinctions of Gala´pagos avifauna. The ap- 1,000 km west of mainland Ecuador in pearance of avian pox–like lesions in both South America. Their isolation and relative domestic chickens and endemic birds on desolation delayed permanent colonization Gala´pagos has heightened these concerns. by humans, and biodiversity remains most- Avian pox is a mild to severe disease of ly intact, with only about 5% of birds that has been reported worldwide in having been lost (Gibbs et al., 1999); this approximately 60 avian species represent- includes none of the 28 breeding land bird ing 20 families. Avian pox is caused by species, 26 of which are endemic. Al- DNA viruses of the family , ge- though 90% of the archipelago was set nus Avipoxvirus, and transmission can oc- aside as a national park in 1959, the resi- cur through introduction into a break in dent human population, along with tour- the skin or, more commonly, when vec- ism, has grown rapidly, and exotics are tored by a biting insect. Disease is most continually being introduced despite in- commonly characterized by cutaneous creasing efforts to exclude them. Because proliferative lesions that harden to thick can have especially severe ef- scabs, but a diptheritic or wet form with fects when introduced to previously isolat- mucosal lesions within the digestive and ed avian populations, as well documented upper-respiratory tracts can occur (Ger-

342 THIEL ET AL.ÐCANARYPOX IN THE GALAÂ PAGOS 343 lach, 1999). The cutaneous form is most a single strain of virus (Ball, 1987); how- commonly observed in passerine birds ever, recombination between coinfecting (Gerlach, 1999). In Gala´pagos, the order viruses may also contribute to the wide di- Passeriformes is represented by eight fam- versity observed between avipoxviruses. ilies with 28 species (Castro and Phillips, The dynamics of multihost pathogens in 1996). Several of these species are severely natural populations is important to under- threatened including the mangrove finch standing general patterns of rapid evolu- (Cactospiza heliobates; total population tion of viruses and their effect on natural approximately 100 individuals), the Flo- populations (Woolhouse et al., 2001; reana mockingbird (Nesomimus trifascia- Cleaveland et al., 2002). In Gala´pagos, tus; approximately 200 individuals), the pox-like symptoms have been described in Espan˜ ola mockingbird (Nesomimus mac- several species of endemic birds, including donaldi; approximately 2,500 individuals), Gala´pagos mockingbirds (N. parvulus par- the medium tree finch (Camarhynchus vulus), Gala´pagos doves, yellow warblers pauper), and the large tree finch (Camar- (D. petechia), and some Gala´pagos finches hynchus psittacula). The diptheritic form (Geospiza spp.) (Jimenez, 2003). Most of the disease is observed most frequently data on the effects of avian pox are from in Psittaciformes, Phasianiformes, and sev- the mockingbirds. eral Columbiformes (Gerlach, 1999). Col- During the 1982–83 El Nin˜ o events, umbiformes present on Gala´pagos include 56% of mockingbirds displaying pox-like a single endemic species, the Gala´pagos lesions died on Genovesa, compared with dove (Zenaida galapagoensis), and the in- 39% of asymptomatic individuals (Curry troduced rock pigeon (Columba livia). and Grant, 1989). In that study, signifi- Thirteen strains of avipoxvirus have cantly more adults were affected than ju- been identified worldwide, and strains vary veniles, partly because the epidemic peak- in virulence and host specificity. Avipoxvi- ed before peak hatching. Prevalence in ruses from endemic forest birds in Hawaii Gala´pagos mockingbirds was higher in (Apapane [Himatione sanguinea] and Ha- nestlings and juveniles than in adults on waiian crow [Corvus hawaiiensis]) include the island of Santa Cruz; a higher resight- two strains that differ significantly from ing rate for young birds without symptoms fowlpox virus by restriction fragment than those with lesions indicated higher length polymorphism genetic analysis (Tri- mortality for infected birds (Vargas, 1987). pathy et al., 2000); their pathogenicity was mild in chickens. Protective immunity also Pox-like lesions were also observed among is strain related, and in trials, birds mockingbirds on Champion (an islet off are often unprotected if challenged with Floreana) during the 1982–83 El Nin˜o different strains (Winterfield and Reed, (Grant et al., 2000). 1985; Sarma and Sharma, 1988; Saini et The first objective of this study was to al., 1990a, b). This indicates that signifi- develop a simple, specific diagnostic test cant antigenic differentiation may exist for avian pox that could be adapted for use among strains, and in part, this may result in the Gala´pagos, where traditional virus from the rapid evolution of avipoxviruses propagation techniques are not economi- by recombination between strains. Repli- cally feasible. The second objective was to cation in avipoxviruses and other viruses in characterize the avipoxviruses that are in- Poxviridae involve genomic intermediates fecting native birds. The availability of comprising many tandem repeats of the large published regions of sequence for entire genome (Moyer and Graves, 1981). the fowlpox and canarypox viruses pro- Recombination, which occurs at extraor- vides the foundation for development of dinarily high frequencies in these viruses, rapid polymerase chain reaction–based is an essential part of replication involving methods for detection of these viruses. 344 JOURNAL OF WILDLIFE DISEASES, VOL. 41, NO. 2, APRIL 2005

METHODS DNA from both viruses (CA3-2F CTAATAGA- Field methods TACTAACGGAGAAG; CA3-2R TTAAATAA- AGAAATGTAAAGAC). Between May and July of 2002 and 2003, wild birds were captured with mistnets near the PCR ampli®cation and sequencing Charles Darwin Research Station on the island PCR amplification with primer set CA3-2 or of Santa Cruz, Galapagos, Ecuador. In January ␮ 2003 and July 2003, birds were captured on the CAX was performed in 50- l volumes of 67 island of Isabela, Galapagos Ecuador. Samples mM Tris-HCl (pH 8.8), 16 mM (NH4)2SO4, 0.01% Tween0, 1.5 mM MgCl2, 2.0 mM each of cutaneous lesions were removed using sterile ␮ scalpels; these were transferred to a plastic vial dNTP, 0.01 mg bovine serum albumen, 0.6 M for each primer, 1 U Taq polymerase (Bioline, and frozen in liquid nitrogen for transport. In ␮ some instances, samples were suspended in Randolph, Massachusetts, USA), and 2 l ex- 95% ethanol and frozen. Any bleeding related tracted DNA (concentration unknown). DNA to sample collection was stopped by applying isolated from fowlpox virus from a chicken in mild pressure with sterile cotton. the United States (kindly provided by D.N. Tri- pathy) was used as a positive control. A touch- DNA extraction down PCR program was used, beginning with an annealing temperature of 50 C and decreas- Samples were frozen in liquid nitrogen, pul- ing 0.5 C every cycle for 14 cycles to a final verized to powder, and incubated at 65 C for annealing temperature of 43 C for an additional at least 6 hr in 250 ␮l Longmire’s lysis buffer 25 cycles. Denaturation was at 94 C, and ex- (0.1 M Tris-HCl, pH 8.0, 0.1 M EDTA, 10 mM tension was 72 C, with a 45-sec hold at each NaCl, 0.5% SDS) with Proteinase-K (final con- temperature in the cycle. Amplicon size was centration, 1.0 mg/mlϪ1). Samples were extract- determined after electrophoresis in a 1.5% aga- ed with phenol/CHCl3/isoamyl alcohol (25:24: rose gel by comparison with markers of known 1), and total DNA was precipitated with etha- size. Amplification with the other primers (see nol and resuspended in 100 ␮l sterile TE (10 previous section) was performed as described mM Tris-HCl, pH 8.0; 10 mM EDTA). for CA3-2, except that the initial annealing temperature varied with the Tm of each primer Primer design pair. For sequencing the 6-kb region of DNA from lesions of samples 502F (designated Gal1) Only one sequence was available for cana- and 100W (designated Gal2) (Table 1), 20 sets rypox virus at the time this project began (6181 of primers were designed that produced 20 am- nucleotides; GenBank D86731), and this was plicons of 350–400 bp, with approximately 50 aligned with the homologous region from fowl- bp of overlap for adjacent amplicons, covering pox virus (complete genome; AF198100) using the 6-kb region. Amplicons were sequenced in ClustalW (Thompson et al., 1994). Overall, the both directions with the same primers used for sequence identity between the two avipoxvirus amplification with the ABI Big Dye protocol on species in this region approximates 70%. The an ABI 377 sequencer. Sequencing of several alignment was visually inspected for regions of amplicons of CA3-2 revealed the presence of divergence, particularly insertions or deletions SpeIorAgeI restriction sites in amplicons from (indels) that would allow rapid differentiation some strains, but not from others; therefore, of PCR products on the basis of size. The in- amplicons of CA3-2 were digested with SpeIor tergenic region between the canarypox virus AgeI in the buffer supplied with the enzyme. genes CA.X and TK (thymidine kinase), desig- The accession number for the sequenced re- nated CAX, was chosen because it provides an gion of Gal1 is AY631870, and for Gal2 it is indel of 52 bp in a region of 426 bp. Primers AY631871 were based on highly conserved sequences within the coding regions of CA.X and TK and Phylogenetic analysis were designed to amplify DNA from both fowl- pox and canarypox viruses (CAXЈF AGATATA- Using ClustalW (Thompson et al., 1994), the GTAGAATTTAGTG; CAXЈR TTCTGCAAGA- sequenced regions of Gal1 and Gal2 were TTTAATATC). The second locus, designated aligned with each other and with the sequence CA3-2, is a region spanning the CA.2 and the of fowlpox virus (Afonso et al., 2000; GenBank CA.3 genes of canarypox viruses. A number of AF19810), and sequences from two canarypox indels in this region led to a predicted differ- viruses described in Tulman et al. (2004; ATCC ence of 18 nucleotides between PCR products VR-111; GenBank AY318871) and Amano et al. from the fowlpox and canarypox viruses. Highly (1999; GenBank D86731). The sequence of the conserved regions within CA.2 and CA.3 pro- culture-adapted fowlpox virus, FP9 (Laidlaw vided the sequence for primers that amplified and Skinner, 2004; GenBank AJ581527), was THIEL ET AL.ÐCANARYPOX IN THE GALAÂ PAGOS 345

TABLE 1. Avipoxviruses from birds in the Gala´pagos Islands.

Strain Species Location Pox type

300C Chicken (Gallus gallus) Santa Cruz Fowlpox 301C Chicken (Gallus gallus) Santa Cruz Fowlpox 302C Chicken (Gallus gallus) Santa Cruz Fowlpox 303C Chicken (Gallus gallus) Santa Cruz Fowlpox 304C Chicken (Gallus gallus) Santa Cruz Fowlpox 305C Chicken (Gallus gallus) Santa Cruz Fowlpox 306C Chicken (Gallus gallus) Santa Cruz Fowlpox 307C Chicken (Gallus gallus) Santa Cruz Fowlpox 308C Chicken (Gallus gallus) Santa Cruz Fowlpox 309C Chicken (Gallus gallus) Santa Cruz Fowlpox 100W Warbler (Dendroica petechia) Santa Cruz Gal2 101W Warbler (Dendroica petechia) Santa Cruz Gal1 102W Warbler (Dendroica petechia) Santa Cruz Gal2 500F Small ground finch (Geospiza fuliginosa) Santa Cruz Gal2 501F Small ground finch (Geospiza fuliginosa) Santa Cruz Gal1 502F Medium ground finch (Geospiza fortis) Santa Cruz Gal1 503F Medium ground finch (Geospiza fortis) Santa Cruz Gal1 504F Medium ground finch (Geospiza fortis) Santa Cruz Gal1 505F Finch (Geospiza sp.) Santa Cruz Gal1 508F Small ground finch (Geospiza fuliginosa) Isabela Gal1 509F Small ground finch (Geospiza fuliginosa) Isabela Gal1 510F Medium ground finch (Geospiza fortis) Isabela Gal1 512F Finch (Geospiza sp.) Unknown Gal1 513F Cactus finch (Geospiza scandens) Isabela Gal1 514F Small ground finch (Geospiza fuliginosa) Isabela Gal2 515F Cactus finch (Geospiza scandens) Santa Cruz Gal2 516F Finch (Geospiza sp.) Santa Cruz Gal1 518F Finch (Geospiza sp.) Santa Cruz Gal1 519F Finch (Geospiza sp.) Santa Cruz Gal1 702M Mockingbird (Nesomimus parvulus) Santa Cruz Gal2 703M Mockingbird (Nesomimus parvulus) Santa Cruz Gal1 identical to the other fowlpox virus sequence (Swofford, 2002), rooted with fowlpox virus. (Afonso et al., 2000; GenBank AF19810) Maximum-likelihood evaluates trees using ex- throughout this region; therefore, it was not in- plicit evolutionary models. MODELTEST 3.06 cluded in the analysis. (Posada and Crandall, 1998) selected the Given the limited nature of our sampling, we TrNϩIϩG evolutionary model as the most like- sought a simple genetic distance-based analysis ly of the 56 possible evolutionary models. The to portray the gross phylogenetic relationships log-likelihood score of the best tree was among the five avipoxvirus sequences. To this 14,404.83620. end, the 5.9-kb alignment was converted to a distance matrix and analyzed via the neighbor- Recombination analysis joining method in PAUP*4.0 (Swofford, 2002). However, the branch leading to fowlpox virus, In Poxviridae, recombination plays an im- which was much longer relative to canarypox portant role in replication, and thus, the fami- virus and canarypox-like Gal1 and Gal2 virus ly’s evolutionary history is not strictly one of sequences, yielded a tree that was biologically association by descent. This has clear conser- untenable (fowlpox virus was a joined sister to vation implications should a bird become in- the Gal strains). Thus, we used a maximum- fected with two viral strains simultaneously. likelihood approach to minimize the problem Thus, we conducted a preliminary recombina- associated with long branches in phylogenetic tion analysis. Specifically, we used statistical analysis (Felsenstein, 1978). A bootstrapped analysis to detect the extent of historical recom- (10,000 replications) maximum-likelihood tree bination among ancestors of the five Poxviridae (with TBR branch swapping) was produced for lineages. This was implemented using the GE- these aligned sequences using PAUP*4.0 NECONV version 1.81 program (an extension 346 JOURNAL OF WILDLIFE DISEASES, VOL. 41, NO. 2, APRIL 2005 of Sawyer, 1989, 2004). This program is a sub- predicted for canarypox virus and larger stitution-based approach that determines than the amplicons produced from the pox whether segments of DNA between two taxa in the alignment are more similar to each other lesions from chickens. than would be expected given their overall level The sequences of the CAX and CA3-2 of similarity. After detecting recombination amplicons from five viruses from chickens events, GENECONV then ranks them accord- from the Gala´pagos and one from the ing to statistical significance and reports where United States were identical to the pub- the recombinatory segment begins and ends within the sequence and its total length. This lished sequence for fowlpox virus at both widely used approach (Millman et al., 2001; loci. Thus, chickens in Gala´pagos were in- Drouin, 2002), is more powerful than other re- fected with an avipoxvirus that is very sim- combination-detection methods in computer ilar, if not identical, to the strain that in- simulations and does not overestimate recom- fects poultry in the United States. In con- bination events (Posada and Crandall, 2001). In our study, pairs of segments of sequences with- trast, the sequences of several CA3-2 am- in the alignment that showed significant recom- plicons from the passerine birds indicated bination are reported as global inner P-values. distinct viruses, but both were very similar Significant recombination between a segment to canarypox virus. One of these ampli- of a sequence from within the alignment and cons, Gal1 (sequenced for 10 strains), con- an unknown taxon outside the alignment, or a taxon within the alignment obscured by other tained restriction sites for SpeI and AgeI, evolutionary processes, are reported here as and the other, Gal2 (sequenced for five global outer P-values. Both are based on 10,000 strains), did not. The Gal1 strain was more permutations, and are corrected for multiple similar to canarypox virus than was Gal2 comparisons. and was also the more prevalent of the two RESULTS strains, particularly in the finches (Table 1). Evidence of avipoxvirus infection was Identi®cation of avipoxviruses restricted to chickens on Santa Cruz that Lesions were collected from 21 endem- were infected with fowlpox virus and the ic passerine birds on the inhabited islands passerine birds that were infected with two of Santa Cruz and Isabela in the Gala´pagos variants of canarypox virus. Because the Islands and from domestic chickens on primers used in this study were designed Santa Cruz (Table 1). The CAX primers using the sequences of fowlpox and cana- amplified a 374-bp fragment of DNA from rypox virus, it is likely that they would not lesions from Gala´pagos chickens (the size amplify DNA from all avipoxvirus strains. predicted from the genome of fowlpox vi- Hence, other avipoxviruses may be present rus) compared to a 426-bp amplicon from in birds in the Gala´pagos. lesions from Gala´pagos finches, yellow Similarity of avipoxviruses from GalaÂpagos warblers, and mockingbirds. The CA3-2 primers amplified a 374-bp fragment of To determine the similarities among the DNA from lesions from Gala´pagos chick- avipoxviruses, we sequenced a 5.9-kb re- ens (the size predicted from the genome gion of a Gal1 (strain 502F) and a Gal2 of fowlpox virus) compared to 392 or 381 (strain 100W) representative strain corre- bp (depending on the avipoxvirus strain) sponding to most of a sequenced region of from lesions from Gala´pagos finches, war- the canarypox virus genome that contains blers, and mockingbirds. Each set of prim- TK gene (Amano et al., 1999; GenBank ers produced an amplicon of the expected D86731). DNA from these two strains size for fowlpox virus from DNA extracted produced single amplicons with all primer from a control fowlpox virus from a U.S. pairs tested, and the sequences of each chicken. The sizes of both the CAX and amplicon were consistent with the pres- CA3-2 amplicons from the lesions from ence of a single virus strain in each spec- the Gala´pagos finches, warblers, and imen. Recently the complete genome se- mockingbirds were comparable to those quence of a slightly different canarypox vi- THIEL ET AL.ÐCANARYPOX IN THE GALAÂ PAGOS 347

FIGURE 1. A maximum-likelihood phylogenetic tree (with TBR branch-swapping) of an alignment of a 5.9-kb region of DNA sequence from five avipoxviruses rooted with fowlpox virus and bootstrapped (10,000 replications); support values are below the branches. Taxon labels are as follows: CanaryVR-111 (Genbank AY31887), Canary (GenBank D86731), Fowlpox (AF198100). rus strain has been published (Tulman et tion using the GENECONV program in- al., 2004; GenBank AY318871), providing dicated 19 significant recombination another strain for comparison to Gal1 and events (global inner fragments) between Gal2. The 5.9-kb sequenced region spans ancestors of four of the five taxa included from within gene CNPV117 to within gene in the analysis (Table 2). Notably, identical CNPV109 (using the nomenclature of the sequence segments or overlapping seg- genes for the complete canarypox virus ge- ments showing evidence of recombination nome [Tulman et al., 2004]). Alignment of occurred between more than one pair of the Gal1 and Gal2 sequences with the two sequences in many cases. More generally, canarypox virus sequences and with fowl- there was statistical evidence of recombi- pox virus confirmed that Gal1 and Gal2 nation at most nucleotide sites along the were most similar to both canarypox virus- 5.9-kb alignment between at least two lin- es. Within this region, sequences for Gal1 eages (Table 2). and Gal2 were 97.6% identical, the two A qualitative analysis of the alignment of published canarypox virus strains were the 5.9-kb region revealed that the iden- 98.7% identical, Gal1 was 97–98% identi- tities among Gal1, Gal2, and canarypox vi- cal to the two canarypox virus strains, and rus were clustered in discrete regions, fur- Gal2 was 95–96% identical to the two can- ther supporting recombination between arypox virus strains. In contrast, sequence strains. In the first 500 nucleotides, Gal1 from fowlpox virus was approximately 70% matched canarypox perfectly, whereas identical to these other strains. The Gal2 differed by about 5%. This was fol- aligned sequences were analyzed using lowed by a region of over 1,500 nucleo- maximum likelihood to produce a phylo- tides in which Gal1 and Gal2 matched genetic tree (Fig. 1). These results were perfectly, and the sequence from canary- fairly congruent with the pairwise compar- pox virus matched them in some regions isons. The canarypox virus, Gal1, and Gal2 but not in others. The next 350 nucleo- strains clustered together and were sepa- tides showed more variability, followed by rated by a long branch leading to fowlpox a similar size region of identity among all virus. three stains. For the remainder of the se- The quantitative analysis of recombina- quence, Gal1 matched canarypox virus, 348 JOURNAL OF WILDLIFE DISEASES, VOL. 41, NO. 2, APRIL 2005

TABLE 2. Putative recombination events between ancestors of avipoxvirus strainsa.

Fragment Sequence 1 Sequence 2 Pb Beginc Endd Lengthe

1 GI Canary CanaryVR-111 0.0206 3605 4402 798 2 GI Canary Gal1 0.0007 464 897 434 3 GI Canary Gal1 0.0128 2591 2973 383 4 GI Canary Gal1 0.0000 3605 4402 798 5 GI Canary Gal2 0.0000 535 897 363 6 GI Canary Gal2 0.0047 2591 2838 248 7 GI Canary Gal2 0.0005 3326 3603 278 8 GI Canary Gal2 0.0002 3605 3940 336 9 GI CanaryVR-111 Gal1 0.0001 1 927 927 10 GI CanaryVR-111 Gal1 0.0000 2476 4431 1956 11 GI CanaryVR-111 Gal1 0.0000 4433 5966 1534 12 GI CanaryVR-111 Gal2 0.0001 535 927 393 13 GI CanaryVR-111 Gal2 0.0126 2521 2838 318 14 GI CanaryVR-111 Gal2 0.0000 3209 3940 732 15 GI CanaryVR-111 Gal2 0.0000 4552 5258 707 16 GI CanaryVR-111 Gal2 0.0020 5343 5684 342 17 GI Gal1 Gal2 0.0000 535 2193 1659 18 GI Gal1 Gal2 0.0001 3209 3940 732 19 GI Gal1 Gal2 0.0013 4552 5258 707 20 GO Fowlpox N/A 0.0000 535 905 371 21 GO Fowlpox N/A 0.0160 2591 2838 248 22 GO Fowlpox N/A 0.0000 3209 3940 732 23 GO Canary N/A 0.0294 2152 2160 9 a Putative recombination events detected using GENECONV version 1.81 (Sawyer, 1989). Nineteen significant (P Ͻ 0.05) recombination events were detected between members of the five-taxon sequence alignment (GI ϭ Global inner fragments), and four significant recombination events were inferred between one taxon within and one unknown taxon outside the alignment or taxa within the alignment obscured by other evolutionary processes (GO ϭ Global outer fragments). b Based on global P value obtained by simulation via 10,000 permutations, corrected for muliple comparisons. c Corresponds to the first nucleotide base of the recombinatory region. d Corresponds to the last nucleotide base of the recombinatory region. e Corresponds to the entire length of the recombinatory region. and Gal2 matched in some regions but not endemic species that have not evolved in others. A region that showed many dif- with or have been previously exposed to ferences among all the strains was in the such pathogens. The losses of endemic 5Ј end of the gene encoding TK birds in Hawaii have been attributed to (CNPV113) (Fig. 2). Many of these differ- the introduction of diseases of domestic ences would result in changes in amino ac- birds and exotic wild birds, including avian ids in the amino-terminal region of the de- pox and malaria (Warner, 1968; van Riper duced proteins. Even the genes of the two et al., 1986, 2002; Tripathy et al., 2000). canarypox virus strains showed differences To date there has been no report of Plas- in the amino acids encoded in this region. modium in Gala´pagos; however, avian pox The mosaic pattern of sequence identity has been prevalent on the islands for de- combined with the many substitutions in cades, infecting both domestic chickens the TK gene indicates that recombination and wild birds. Viruses in Poxviridae, in- between avipoxvirus strains, combined cluding avipoxviruses often cross species with mutation, led to the variants of can- barriers (Boosinger et al., 1982; Reed et arypox-like viruses that are now prevalent al., 2004). The presence of chickens on the on Santa Cruz and Isabela. inhabited islands of Gala´pagos, therefore, DISCUSSION has led to local concern that an avipoxvirus The introduction of domestic animal of domestic birds could spread to the en- diseases to archipelagos poses a threat to demic wild bird populations. THIEL ET AL.ÐCANARYPOX IN THE GALAÂ PAGOS 349

FIGURE 2. Alignment of thymidine kinase. The deduced amino acids for the putative thymidine kinase genes of five avipoxviruses were aligned using ClustalW. Black highlights indicate residues that show some variability among all five strains. Grey highlights indicate residues that differ only in fowlpox virus. Canary1 from GenBank D86731; Canary2 from Genbank AY31887; Fowl from Genbank AF198100.

Restriction fragment length polymor- poxviruses may have been introduced by phisms for the CA3-2 amplicon, as well as migrant birds from other islands. Because sequencing of a 5.9-kb region of the ge- these viruses are mechanically vectored, it nome, indicate that endemic passerine can be transferred to a new host by any birds, including finches, warblers, and number of biting insects, as well as by in- mockingbirds, were infected with one of troduction of virions through any break in two closely related variants of canarypox the skin. Vector communities differ be- virus, but that none were infected with tween uninhabited islands and the human- fowlpox virus. In contrast, chickens from inhabited islands where biting insects that Santa Cruz were infected with fowlpox vi- may require fresh water are found (e.g., rus but the canarypox-like virus strains the Culex quinquefasciatus and were not detected. These results indicated the blackfly Simulium bipunctatum) (Peck that on Santa Cruz and Isabela, there was et al., 1998). On islands without fresh wa- no evidence of avipoxvirus transmission ter, the mosquito Ochlerotatus taeniorhyn- between endemic and domestic birds. chus and a number of hippoboscid para- However, the similarity of the Gal1 and sitic flies and ceratopogonid biting midges Gal2 strains to canarypox virus indicated are more common. that these viruses originated from either The of all characterized pox- wild or domestic passerine birds, through viruses comprise a single chromosome of human involvement or natural migrations. linear double-stranded DNA that has telo- It is unknown whether the canarypox-like meric ends with covalently closed terminal strains detected in this study represent di- hairpin loops. The virus encodes all the vergence from a single ancestor by a com- proteins required for DNA replication, bination of mutation and recombination, which occurs in the cytoplasm of the host or whether they are the result of multiple cell. Replication of the genome occurs introductions. It will be interesting to through intermediates called concatamers, characterize avipoxvirus strains from birds comprising many tandem repeats of the on the other islands of the Gala´pagos, par- entire genome (Moyer and Graves, 1981). ticularly the uninhabited islands where avi- Recombination, which occurs at extraor- 350 JOURNAL OF WILDLIFE DISEASES, VOL. 41, NO. 2, APRIL 2005 dinarily high frequency, is an essential part strains, Gal1, and Gal2 portrayed here in- of concatamer formation and replication dicates that the two avipoxvirus strains (Ball, 1987). The virus-encoded DNA identified from endemic, wild, Gala´pagos polymerase of virus mediates re- are close relatives within the combination; mutants lacking this poly- canarypox lineage. However, because no merase do not recombine DNA (Mer- strains from native, wild, mainland South chlinsky, 1989; Willer et al., 1999, 2000). American birds were included in this anal- Consistent with the role for recombination ysis, it is impossible to determine the na- in the replication of avipoxviruses, the re- ture and number of avipoxvirus introduc- combination pathway in vaccinia virus can tions into the Gala´pagos Archipelago. Nev- very efficiently recombine pairs of linear ertheless, the Gal1 and Gal2 strains are molecules and requires only 12–20 bp of much more similar to other passerine avi- homology (Willer et al., 2000; Yao and poxviruses (e.g., canarypox virus) than they Evans, 2001). are to fowlpox virus. This phylogenetic re- Recombination is thought to serve a construction should be accepted with cau- number of possible functions in viruses. tion because of its narrow scope and be- Because there is no known primase, re- cause recombination can cause consider- combination may play a role in the prim- able error in tree estimation. There is clear ing of DNA replication. It can function to evidence of recombination in the sequenc- repair double-stranded breaks in DNA, es of the 5.9-kb region. On Gala´pagos, the and it may provide a mechanism for the sympatry of fowlpox virus in the intro- acquisition of new genes from a coinfect- duced chickens and the canary poxvirus ing virus or from the host cell (Yao and variants we describe here in endemic birds Evans, 2001). The similarity of many viral presents an opportunity for further recom- genes with mammalian genes provides binants of unknown effect. strong evidence for the acquisition of host The TK gene showed the greatest di- genes by poxviruses, although the mecha- vergence of any gene in the 5.9-kb region, nism is not understood (Yao and Evans, even in the canarypox-like viruses reported 2001). Intermolecular recombination be- in this study. This viral enzyme is part of tween the genomes of different viruses has the salvage pathway that allows the poxvi- been implicated in the formation of new ruses to phosphorylate nucleotides for recombinant poxvirus strains. Malignant DNA synthesis. Although the gene is not rabbit fibroma virus, a lethal tumorigenic essential, it is present in the genomes of poxvirus of rabbits, resulted from recom- all poxviruses sequenced to date except bination between Shope fibroma virus, Molluscum contagiosum (Gubser et al., which induces benign tumors in rabbits, 2004). The difference in the amino acid and myxoma virus, which causes myxo- sequences of the enzymes from very close- matosis (Block et al., 1985; Upton et al., ly related strains indicates that it may 1988). In another instance, the genome evolve rapidly, and hence it may serve as structures of one capripoxvirus isolate in- a marker for identification of different avi- dicated that the progenitor of this strains poxvirus strains and for measuring the rate resulted from recombination between the of viral evolution on different islands in genomes of two other capripoxvirus strains the Gala´pagos. (Gershon and Black, 1988; Gershon et al., Avipoxvirus infection can significantly 1989). Thus, recombination between dif- increase mortality of Galapagos mocking- ferent avipoxvirus strains may provide a birds (Curry and Grant, 1989; Vargas, powerful mechanism for rapid evolution of 1987). Presumably, it can have similar con- poxviruses in wild animals. sequences for the other susceptible Gala- The gross phylogenic relationships pagos endemics in which it occurs, al- among fowlpox virus, two canarypox virus though these have not been measured. THIEL ET AL.ÐCANARYPOX IN THE GALAÂ PAGOS 351

Management plans for small populations Shope fibroma virus and myxoma virus. Virology further threatened by pathogens requires 140: 113–124. BOOSINGER, T. R., R. W. WINTERFIELD,D.S.FELD- characterizing the pathogens and an un- MAN, AND A. S. DHILLON. 1982. Psittacine pox derstanding of how such pathogens are in- virus: Virus isolation and identification, transmis- troduced. This work indicates that the two sion, and cross-challenge studies in parrots and lineages of avipoxvirus described in en- chickens. Avian Diseases 26: 437–444. demic passerines did not arrive through CASTRO, I., AND A. PHILLIPS. 1996. A guide to the birds of the Gala´pagos Islands. Princeton Uni- introduced chickens. The work also indi- versity Press, Princeton, New Jersey, pp. 91–131. cates, however, that significant recombi- CHRISTIE, D. M., R. A. DUNCAN,A.R.MCBIRNEY, nation continues to occur among the M. A. RICHARDS,W.M.WHITE,K.S.HARPP, strains in the endemic birds, which could AND C. G. FOX. 1992. Drowned islands down- continue to generate new forms with pos- stream from the Gala´pagos hotspot imply ex- tended speciation times. Nature 355: 246–248. sible differences in pathogenicity and host CLEAVELAND, S., G. R. HESS,A.P.DOBSON,M.K. range. Efforts to understand the pathoge- LAURENSON,H.I.HCCALLUM,M.G.ROBERTS, nicity of the two extant strains, to under- AND R. WOODROFFE. 2002. The role of patho- stand predominant routes of transmission, gens in biological conservation. In The ecology especially with regard to vectors, and to of wildlife diseases, P. J. Hudon, A. Rizzoli, B. T. Grenfell, H. Heesterbeek, and A. P. Dobson continue to characterize avipoxviruses (eds.). Oxford University Press, Oxford, UK, pp. from both endemic and exotic avian spe- 139–150. cies are recommended. CURRY, R. L., AND P. R. GRANT. 1989. Demography of the cooperatively breeding Gala´pagos mock- ACKNOWLEDGMENTS ingbird, Nesomimus parvulus, in a climatically variable environment. Journal of Animal Ecology This work was supported by the University 58: 441–463. of Missouri–St. Louis, the Saint Louis Zoo, and DROUIN, G. 2002. Characterization of the gene con- the Des Lee Professorship in Zoological Stud- versions between the multigene family members ies, which unites those two institutions. We of the yeast genome. Journal of Molecular Evo- thank Mary Duncan at the St. Louis Zoo for lution 55: 14–23. first identifying avipoxviruses in endemic birds FELSENSTEIN, J. 1978. Cases in which parsimony in the Gala´pagos by histological techniques. and compatibility methods will be positively mis- The Charles Darwin Research Station and the leading. Systematic Zoology 27: 401–410. Gala´pagos National Park provided logistical GERLACH, H. 1999. Viruses. In Avian medicine: support. We especially thank David Wieden- Principles and application, B. Ritchie, G. Harri- feld, Poly Robayo, Susana Cardenas, and How- son, and L. Harrison (eds.). Wingers Publishing, ard Snell, for their support in Gala´pagos. We Lake Worth, Florida, pp. 862–948. thank Kelly Halbert for excellent technical as- GERSHON,P.D.,AND D. N. BLACK. 1988. A com- sistance in the lab and D.N. Tripathy for pro- parison of the genomes of capripoxvirus isolates viding a sample of DNA from a U.S. isolate of of sheep, goats, and cattle. Virology 164: 341– fowlpox virus. 349. LITERATURE CITED ,R.P.KITCHING,J.M.HAMMOND, AND D. N. BLACK. 1989. Poxvirus genetic recombina- AFONSO, C. L., E. R. TULMAN,Z.LU,L.ZSAK,G.F. tion during natural virus transmission. Journal of KUTISH, AND D. L. ROCK. 2000. The genome General Virology 70: 485–489. of fowlpox virus. Journal of Virology 74: 3815– GIBBS, J. P., H. L. SNELL, AND C. E. CAUSTON. 1999. 3831. Effective monitoring for adaptive wildlife man- AMANO, H., S. MORIKAWA,H.SHIMIZU,I.SHOJI,D. agement: Lessons from the Gala´pagos Islands. KUROSAWA,Y.MATSUURA,T.MIYAMURA, AND Y. Journal of Wildlife Management 63: 1055–1065. UEDA. 1999. Identification of the canarypox vi- GRANT, P. R., R. L. CURRY, AND B. R. GRANT. 2000. rus thymidine kinase gene and insertion of for- A remnant population of the Floreana mocking- eign genes. Virology 256: 280–290. bird on Champion island, Gala´pagos. Biology BALL, L. A. 1987. High-frequency homologous re- Conservation 92: 285–290. combination in vaccinia virus DNA. Journal of GUBSER, C., S. HUE,P.KELLAM, AND G. L. SMITH. Virology 61: 1788–1795 2004. Poxvirus genomes: A phylogenetic analy- BLOCK, W., C. UPTON, AND G. MCFADDEN. 1985. sis. Journal of General Virology 85: 105–117. Tumorigenic poxviruses: Genomic organization JIMENEZ, U. G. 2003. Estudio de la viruela aviar: of malignant rabbit virus, a recombinant between Especies silvestres afectadas, comparacion entre 352 JOURNAL OF WILDLIFE DISEASES, VOL. 41, NO. 2, APRIL 2005

epocas y anteriores estudios. Isla Santa Cruz, Ga- gene conversion. Molecular Biology and Evolu- lapagos, Ecuador. In XXVII Jornadas Ecuatori- tion 6: 526–538. anas de Biologia ‘‘Pedro Nunez Lucio.’’ Sociedad . 2004. GENECONV. http://www.math. Ecuatoriana de Biologia. Escuela de Biologia de wustl.edu/ϳsawyer. Accessed April 2005. la Universidad Central del Ecuador, E. Ceron SWOFFORD, D. L. 2002. PAUP*. Phylogenetic Anal- and C. Reyes (eds.). Quito, Ecuador, pp. 86–87. ysis Using Parsimony (*and Other Methods). LAIDLAW, S. M., AND M. A. SKINNER. 2004. Com- Version 4. Sinauer Associates, Sunderland, Mas- parison of the genome sequence of FP9, an at- sachusetts. tenuated, tissue culture-adapted European strain THOMPSON, J. D., D. G. HIGGINS, AND T. J. GIBSON. of Fowlpox virus, with those of virulent American 1994. CLUSTALW: Improving the sensitivity of and European viruses. Journal of General Virol- progressive multiple sequence alignment ogy 85: 305–322. through sequence weighting, position specific MERCHLINSKY, M. 1989. Intramolecular homolo- gap penalties, and weight matrix choice. Nucleic gous recombination in cells infected with tem- Acids Research 22: 4673–4680. perature-sensitive mutants of vaccinia virus. TRIPATHY,D.N.AND W. M. SCHNITZLEIN. 1999. Journal of Virology 63: 2030–2035. Fowlpox virus (Poxviridae). In Encyclopedia of MILLMAN, K. L., S. TAVARE´ , AND D. DEAN. 2001. virology, A. Granoff and R. G. Webster (eds.), Recombination in the ompA gene but not the Academic Press, London, UK, pp. 576–582. omcB gene of Chlamydia contributes to serovar- , AND W. M. REED. 1997. Pox. In Diseases specific differences in tissue tropism, immune of poultry, 10th ed., B. W. Calnek, H. J. Barnes, surveillance, and persistence of an organism. C. W. Beard, L. R. McDougald, and Y. M. Saif Journal of Bacteriology 183: 5997–6008. (eds.). Iowa State University Press, Ames, Iowa, MOYER,R.W.,AND R. L. GRAVES. 1981. The mech- pp. 643–659. anism of cytoplasmic orthopoxvirus DNA repli- , AND . 1998. Pox. In A manual for the cation. Cell 27: 391–401. isolation and identification of avian pathogens, PECK, S. B., J. HERATY,B.LANDRY, AND B. J. SIN- 4th edition. D. E. Swayne, J. R. Glisson, M. W. CLAIR. 1998. Introduced insect fauna of an oce- Jackwood, J. E. Pearson, and W. M. Reed (eds.). anic archipelago: The Gala´pagos Islands, Ecua- American Association of Avian Pathologists, New dor. American Entomologist 44: 218–237. Bolton Center, Pennsylvania, Pennsylvania, pp. POSADA, D., AND K. A. CRANDALL. 1998. MODEL- 137–140. TEST: Testing the model of DNA substitution. ,W.M.SCHNITZLEIN,P.J.MORRIS,D.L. Bioinformatics 14: 817–818. JANSSEN,J.K.ZUBA,G.MASSEY, AND C. T. AT- , AND . 2001. Evaluation of methods KINSON. 2000. Characterization of poxviruses for detecting recombination from DNA sequenc- from forest birds in Hawaii. Journal of Wildlife es: Computer simulations. Proceedings of the Diseases 36: 225–230. National Academy of Sciences, USA 98: 12757– TULMAN, E. R., C. L. AFONSO,Z.LU,L.ZSAK,G.F. 13762. KUTISH, AND D. L. ROCK. 2004. The genome REED, K. D., J. W. MELSKI,M.B.GRAHAM,R.L. of canarypox virus. Journal of Virology 78: 353– REGNERY,M.J.SOTIR,M.V.WEGNER,J.J.KAZ- 366. MIERCZAK,E.J.STRATMAN,Y.LI,J.A.FAIRLEY, UPTON, C., J. L. MACEN,R.A.MARANCHUK,A.M. G. R. SWAIN,V.A.OLSON,E.K.SARGENT,S.C. DELANGE, AND G. MCFADDEN. 1988. Tumori- KEHL,M.A.FRACE,R.KLINE,S.L.FOLDY,J. genic poxviruses: Fine analysis of the recombi- P. D AVIS, AND I. K. DAMON. 2004. The detec- nation junctions in malignant rabbit fibroma vi- tion of monkeypox in humans in the Western rus, a recombinant between Shope fibroma virus Hemisphere. New England Journal of Medicine and myxoma virus. Virology 166: 229–239. 350: 342–350. VAN RIPER, C., III, S. G. VAN RIPER, AND W. R. SAINI, S. S., N. K. MAITI, AND S. N. SHARMA. 1990a. HANSEN. 2002. Epizootiology and effect of avi- Immune response of chicks to oral vaccination an pox on Hawaiian forest birds. Auk 119: 929– with combined extracellular and intracellular 942. fowl pox viruses. Animal Health and Production , ,M.L.GOFF, AND M. LAIRD. 1986. 22: 165–169. The epizootiology and ecological significance of ,S.S.SODHI,N.K.MAIKI, AND S. N. SHAR- malaria in Hawaiian land birds. Ecological MA. 1990b. Immune response of chicks to oral Monographs 56: 327–344. vaccination against Newcastle disease and fowl VARGAS, H. 1987. Frequency and effect of pox-like pox. Comparative Immunology, Microbiology lesions in Gala´pagos mockingbirds. Journal of and Infectious Diseases 13: 1–6. Field Ornithology 58: 101–102. SARMA, D. K., AND S. N. SHARMA. 1988. Compara- WARNER, R. E. 1968. The role of introduced dis- tive immunity of fowl pox virus . Zen- eases in the extinction of the endemic Hawaiian tralblatt fur veterinarmedizin Reihe B 35: 19–23. avifauna. Condor 70: 101–120. SAWYER, S. A. 1989. Statistical tools for detecting WHITE, W. M., A. R. MCBIRNEY, AND R. A. DUNCAN. THIEL ET AL.ÐCANARYPOX IN THE GALAÂ PAGOS 353

1993. Petrology and geochemistry of the Gala´- 2000. In vitro concatemer formation catalyzed pagos Islands: Portrait of a pathological mantle by vaccinia virus DNA polymerase. Virology 278: plume. Journal of Geophysical Research 98: 562–569. 19533–19563. WINTERFIELD,R.W.,AND W. R EED. 1985. Avian WIKELSKI, M., J. FOUFOPOULOS,H.VARGAS, AND pox infection and immunity with quail, psitta- H. SNELL. 2004. Gala´pagos birds and diseas- cine, fowl and viruses. Poultry Sci- es: Invasive pathogens as threats for island spe- ence 64: 65–70. cies. Ecology and Society 9: 5; http://www. WOOLHOUSE, M. E. J., L. H. TAYLOR, AND D. T. HAYDON. 2001. The population biology of mul- ecologyandsociety.org/vol9/iss1/art5. Accessed ti-host pathogens. Science 292: 1109–1112. April 2005. YAO, X. D., AND D. H. EVANS. 2001. Effects of DNA WILLER, D. O., M. J. MANN,W.ZHANG, AND D. H. structure and homology length on vaccinia virus EVANS. 1999. Vaccinia virus DNA polymerase recombination. Journal of Virology 75: 6923– promotes DNA pairing and strand-transfer re- 6932. actions. Virology 257: 511–523. ,X.D.YAO,M.J.MANN, AND D. H. EVANS. Received for Publication 21 July 2004.